It has recently been hypothesized that invariant acoustic cues for place of articulation can be derived from the short-time spectrum at consonantal release. These invariant properties can be characterized in terms of the gross shape of the spectrum and theoretically provide the listener with perceptual invariance independent of context-dependent effects. This research will test this model of speech perception by assessing whether invariant acoustic cues are in fact perceived independent of context effects of absolute onset frequency, vowel context, and syllable position. Further, we will attempt to delineate the nature of the perceptual mechanism for interpreting onsets by investigating the window or frame of analysis for onset spectra and the notion of spectral discontinuity. To test these questions, a series of perception experiments using synthetic stimuli varying in the acoustic cues provided will be conducted. These experiments will use various paradigms including categorization tests, selective adaptation, dichotic listening, and judgements concerning specific attributes of the stimuli.